1. Field of the Invention
The present invention relates to a multi-focus camera apparatus that photographs the same region by a plurality of cameras having different focal lengths, and to an image processing method and an image processing program used therein.
2. Description of the Related Art
In images photographed for broadcasting or monitoring, methods for protecting personal information by mosaic processing on target objects such as people, pets like dogs and cats, and vehicles have been used. In addition, techniques using image data encryption to prevent image viewing by a third person have been known.
In methods using such mosaic processing or the like, detection of regions representing target objects in photographed images is especially necessary.
As a method of detecting a target object in an image, various kinds of known methods have been used. For example, International Patent Publication No. WO02/073535 describes a method of detecting a region of a target object representing visual information for car navigation such as a road sign and an address in images photographed by a plurality of cameras having different focal lengths installed in a vehicle, through pattern recognition according to an artificial intelligence algorithm.
However, in the above-described method according to pattern recognition, the detection performance unfortunately tends to deteriorate due to a change in a photography condition and a movement of a subject, for example. Consequently, a protection target object may happen to be displayed without mosaic processing thereon, due to failure of detection of a portion or even a whole of the target object. Therefore, appropriate protection of personal information is not achieved in some cases.